Documente Academic
Documente Profesional
Documente Cultură
The goal of this study was to train, validate, select and evaluate artificial neural networks (ANN) to
predict the individual volume of wood in eucalyptus stand, based on the diameter at breast height
(DBH) and DBH with the total height (Ht). Data was obtained from a plantation of Eucalyptus urophylla
ST Blake of seven years of age, located in the state of Gois, Brazil. Sixteen plots were randomly set in
this area, from which the variables diameter, total height and volume were accounted. The volume of all
the trees in each plot was measured by the Smalian method; afterwards, the data were checked for
normality using the Shapiro-Wilk test. Sequentially, perceptron network settings (ANN1 = DBH and Ht;
and ANN2 = DBH) were trained using sigmoid activation functions and resilient propagation (Rprop)
algorithm. In addition, a root-mean-square error (RMSE) of less than 1% was adopted as stopping
criterion or when this rose again. The selected ANNs presented low variation among the task-specific
training indices, selection and evaluation, showing correlation ( ) between predicted and observed
volume (0.9945 and 0.9898), and RMSE from 1.75 and 2.22%, respectively. The Shapiro-Wilk test
highlighted non-normality of data distribution; hence, various selected ANNs were subjected to the
Kruskal-Wallis test
for validation, as well as for comparison with each other and sequentially
submitted to the overall group difference test. The test
demonstrated that both ANNs were able to
predict tree volume; leading to the conclusion that multilayer perceptron neural networks (MLPNNs),
using just one neuron input- the diameter, are as precise and accurate as networks using two neuronsthe diameter and height, in order to predict individual volume of E. urophylla.
Key words: Eucalyptus, Brazil, volumetry, forest inventory, neural networks.
INTRODUCTION
Planted forests cover about 264 million hectares,
comprising nearly 7% of total forest area, with the largest
part located in China, India and the United States (61%).
Miguel et al.
1915
1916
Figure 1. Study area location and spatial distribution of sampled plots. Source: Google Earth image; Natural Earth
(naturalearthdata.com); IBGE (Brazilian Institute of Geography and Statistics).
t 2 .Cv% 2
n
E2
(1)
Miguel et al.
Yk (Vk )
(2)
= Activation function;
1917
Vk xn .wn
(3)
o
The networks were also comprised of a single hidden layer
architecture. According to Esquerre (2002), most of the time, the
networks require a single hidden layer to solve nonlinearly
separable problems. The number of neurons in this layer has been
optimized by the Intelligent Problem Solver (IPS) tool of the
Statistica 7 software (Statsoft, 2007); and as activation function, the
sigmoidal was used.
Sigmoid activation function is quite common in developing
artificial neural networks, and in addition to a well-built network
architecture, it can bring closer any continuous function with great
precision (Ismailov, 2014); this function is mathematically expressed
by:
(v )
Where,
v
n
1 exp u
100
RMSE(%)
v
(5)
(4)
1918
Figure 2. Normal distribution probability (A) and diameter distribution in a stand of E. urophylla S. T. Blake sampled in the state of Gois,
Brazil (B).
Table 1. Frequency distribution of the optimal number ( = 0.05; 5%) of measured tree for training of the ANNs, by
diameter and height classes for Eucalyptus urophylla in the state of Gois, Brazil.
Class center
Ht (m)
12.5
15.5
18.5
21.5
24.5
27.5
30.5
Total
6.25
2
3
3
2
10
8.75
1
2
3
4
4
14
18.75
5
6
5
5
4
25
21.25
2
3
3
3
12
Total
3
12
24
30
31
22
9
131
Miguel et al.
ANN 1
1919
ANN 2
Figure 3. Architecture of artificial neural networks (ANN) for prediction of individual total wood volume of forest stands of E. urophylla, in
function of the DBH and Ht (ANN1) and DBH (ANN2), in the state of Gois, Brazil.
Table 2. Characteristics and performance statistics of the artificial neural networks (ANN) selected to
estimate the volume of E. urophylla trees in the city of Rio Verde, state of Gois, Brazil.
Network
Predicting
variables
ANN1
ANN2
Neurons by layer
Input hidden output
2
5
1
1
7
1
LT
0.08
0.10
LS
0.08
0.12
Adjustment
LA
RMSE (%)
0.08
0.994
1.75
0.11
0.989
2.22
IT = Levels of training (network acquisition), IS = levels of stop selection (training stop), IA = levels of assessment
(trained network quality), = correlation between observed and predicted volume, RMSE% = root-mean-square
error in percentage.
ANN training
Adjustment and accuracy statistics of both selected
networks to predict individual wood volume of E.
urophylla trees were satisfactory. The selected ANNs
showed low variation between the levels of training,
selection and evaluation, which consist of ideal results
that show training stability (Binoti et al., 2013). The
correlation between observed and predict volumes ( )
was 0.9945 and 0.9898, with root-mean-square error in
percentage (RMSE%) between 1.75 and 2.22%,
respectively (Table 2). Thus, the use of the ANNs has
1920
Figure 4. Behavior in the training of ANN1 (DBH and Ht) and ANN2 (DBH) in predicting individual wood volume (A), residual distribution (B)
and the error frequency histogram (C) for stands of E. urophylla in the city of Rio Verde, state of Gois, Brazil.
Miguel et al.
1921
Table 3. Mean, minimum and maximum values, as well as real and estimated values of individual wood
volume, estimated by both network categories, and validation statistics of E. urophylla in Rio Verde,
state of Gois, Brazil.
Variable
Minimum
Mean
Maximum
Da (%)
0.0196
0.0192
0.0197
0.2682
0.2703
0.2703
0.6138
0.6107
0.5725
-0.75
-1.75
11.05
27.51
66.62
66.62
ns
ns
0.6107
0.5725
-1.00
16.50
66.62
ns
Actual volume
ANN1
ANN2
D.cal
D.ref
Result
( Dcal = Value calculated using Kruskal-Wallis test, ( Dref = reference value according to the Kruskal-Wallis
test, Da (%) = aggregated difference, ns = not significant at 5%.
Figure 5. Behavior in validations of ANN1 (DAP and Ht) and ANN2 (DAP) in predicting individual wood volume (A), residual
distribution (B) and error frequency histogram (C) for forest stands of E. urophylla in the city of Rio Verde, state of Gois,
Brazil.
1922
Conclusions
Artificial neural networks of the type multilayer perceptron
using the diameter and the total tree height as predictor
variables are accurate in the estimate of individual wood
volume of E. urophylla trees and are not statistically
different from the true wood volume derived from rigorous
tree scaling process. Settings of multilayer perceptron
networks with a single neuron in the input layer, the
diameter, are as precise and accurate as networks using
two neurons in the same layer, the diameter and height.
Conflict of Interests
The authors have not declared any conflict of interests.
REFERENCES
Azevedo TL, Mello AA, Ferreira RA, Sanquetta CR, Nakajima NY
(2011). Equaes hipsomtricas e volumtricas para um povoamento
de Eucalyptus sp. localizado na FLONA do Ibura, Sergipe. Rev. Bras.
Cienc. Agrrias 6(1):105-112.
Binoti MLMS, Binoti DHB, Leite HG (2013). Aplicao de redes neurais
Miguel et al.
Miguel EP, Leal FA, Ono HA, Leal UAS (2014). Modelagem na predio
do volume individual em plantio de Eucalyptus urograndis. Rev. Bras.
Biometria 32(4):584-598.
Miguel EP, Rezende AV, Leal FA, Matricardi, EAT, Vale Miguel EP,
Rezende AV, Leal FA, Matricardi EAT, do Vale AT, Pereira RS
(2015). Redes neurais artificiais para a modelagem do volume de
madeira e biomassa do cerrado com dados de satlite.Pesqui.
Agropecu. Bras. 50(9):829-839.
Pereira W, Tanaka OK (1990). Estatstica: Conceitos bsicos. 2. ed.
So Paulo: Makron Books 371 p.
Prez-Cruzado C, Merino A, Rodrguez-Soalleiro R (2011). A
management tool for estimating bioenergy production and carbon
sequestration in Eucalyptus globulus and Eucalyptus nitens grown as
short rotation woody crops in north-west Spain. Biomass Bioenergy
35(7):2839-2851.
Riedmiller M, Braun H (1993). A direct adaptive method for faster
backpropagation learning: Therporp algorithm. Int. Join. Conf. Neural
Netw. 1(7):586-591.
Sanquetta CR, Behling A, Dala Corte AP, Ruza MS, Simon A, So Jos
JFB (2014). Relao hipsomtrica em inventrios pr-corte em
povoamentos de Acacia mearnsii De Wild. Cientfica 42(1):80-90.
Scolforo JRS, Thiersch CR (2004). Biometria florestal: Medio,
volumetria e gravimetria. Lavras: UFLA/FAEPE 285 p.
Shapiro SS, Wilk MB (1965). An analysis of variance test for normality
(complete sample). Biometrika 52(3):591-611.
Shiblee MD, Chandra B, Kalra PK (2010). Learning of geometric mean
neuron model using resilient propagation algorithm. Expert Syst.
Appl. 37(12):7449-7455.
Silva IN, Spatti DH, Flauzino RA (2010). Redes Neurais Artificiais para
engenharia e cincias aplicadas. Artliber. 397 p.
Silva MLM, Binoti DHB, Gleriani JM, Leite HG (2009). Ajuste do modelo
de Schumacher e Hall e aplicao de redes neurais artificiais para
estimar volume de rvores de eucalipto. Rev. Arvore 33(6):11331139.
1923